Vol. 8, No. 1, January 1969
CIRCULAR DICHROISM O F COBALT(II1) COMPLEXES llci
(f) The iron to azopyridine charge-transfer band lies a t an exceptionally low energy. Other things being equal, the difference in energy between the metal t Z g orbitals and the empty ligand orbitals into which they are delocalized (which are higher in energy than the metal tZg orbitals) will be smaller in this case than in the imine ligands previously investigated. This situation will favor extensive delocalization. We might anticipate that the extent of T delocalization will increase with decreasing energy of the iron to ligand chargetransfer band. All of the items are indicative of a remarkably strong
interaction between the iron atom and azopyridine. The items (b) and (d) seem especially indicative of a T-delocalization mechanism playing an important role in this interaction68 If item (d) is correctly interpreted, we may infer that the iron atom should act as an electron-donating substituent in determining, for example, the substitution pattern of the ligand. Experiments to see if this is so are being undertaken. Acknowledgment.-D. A. B. is grateful to the Science Research Council (Great Britain) for a research studentship.
CONTRIBUTION FROM THE DEPARTMENT O F CHEMISTRY, UNIVERSITY OF PITTSBURGH, PITTSBURGH, PENNSYLVANIA 15213
Circular Dichroism of Cobalt(II1) Complexes with Ethylenediamine-N,N’-diacetic Acid and a Carbonate, Oxalate, or Malonate Anion1 BY CAROLYN W. VAN S d U N
AND
BODIE E. DOUGLAS
Received July 31, 1968 Preparation and resolution procedures are reported for a series of cobalt(II1) complexes with the tetradentate ligand ethylenediamine-N,N’-diacetic acid (EDDA) and a carbonate, oxalate, or malonate anion. Proton magnetic resonance was used to establish the trans coordination of the oxygens of EDDA in isomers studied. Absolute configurations of the new complexes are assigned by comparison of their CD curves with that of (+)jss-trans-Co(EDDA)en+, for which the absolute configuration is known. Rotational strengths of the d-d transitions were found to be determined largely by the EDDA and to be quite insensitive to the bidentate ligand used.
Introduction While the relationship between the absolute configuration of cobalt(II1) complexes containing chelate rings and the sign of the Cotton effects of their d-d transitions has received considerable attention, the magnitude of the rotational strengths of these transitions has not been studied systematically. Reasons for variations in intensities of circular dichroism (CD) bands within and between series of related complexes are not known. With the aim of assessing the influence of ring size on rotational strength, a new series of cobalt(II1) complexes has been prepared with each complex containing the tetradentate ligand ethylenediamine-N,N’-diacetate (EDDA) and a carbonate or a dicarboxylate anion. The anions used were carbonate, oxalate (ox), and malonate (mal) giving four-, five-, and six-membered chelate rings, respectively. The geometrical isomers studied were those with the coordinated oxygens of the EDDA trans to one another. For the previously studied complexes containing these bidentate anions, no consistent relationship between ring size and rotational strength has been found. (1) This work was supported, in part, by a research grant (GM 10829-10) from the Division of General Medical Studies, Public Health Service. (2) S. Kirchner, Coovd. Chem. Rev., 2, 461 (1967); J. Fujita and Y. Shimura in “Spectroscopy and Structure of Metal Chelate Compounds,” K. Nakamoto and P. J. McCarthy, Ed., John Wiley & Sons, Inc., New York, N. Y . , 1968, p 156.
For the complex ions Co(en) (ox)Z- and Co(en) (mal)z-, it was found that the rotational strength in the visible region for the malonato complex was approximately twice that for the oxalato ~ o m p l e x . ~On the other hand, for the ions Co(g1y) (ox)zz- and Co(g1y) (ma1)Z2-, the rotational strength in the visible region for the oxalato complex is about twice that for the malonato ~omplex.~ For the series Co(en)zX+, where X = C03*-, ox, or mal, the carbonato complex gave the highest rotational strength and the malonato complex gave the lowest, but it was thought that complete resolution was not achieved for the malonato ~ o m p l e x . ~ More circular dichroism data are needed to determine whether it is ring size or some other property of these anions which is important for determining the rotational strengths of the d-d transitions of these complexes. Experimental Section The optically active isomers of the new complexes are characterized by (+) or (-), the sign of the lowest energy CD maximum. Ethylenediamine-N,N’-diacetic acid was obtained from K & K Laboratories, Inc., or from Pfaltz and Bauer, Inc., and used without further purification. (3) B. E. Douglas, R. A. Haines, and J. G. Brushmiller, Inovg. Chem., 2, 1194 (1963). (4) K. Yamasaki, J. Hidaka, and Y . Shimura, private communication. (5) R. A. Haines, Ph.D. Thesis, University of Pittsburgh, 1964. Resolution being checked.
118 CAROLYN U'. VAN SAUNAND RODIE E. DOUGLAS Preparation of Potassium trans-Ethylenediamine-N,N'-diacetato(carbonato)cobaltate(III) Tetrahydrate, K[trans-Co(EDDA)C03].4HzO.-A solution of CoC12.6Hz0 (11.9 g, 0.05 mol) in water (10 ml) and 30y0 H202 (10 ml) was added dropwise to a cold slurry of KHC03 (25 g, 0.25 mol) in water (20 ml). The mixture was cooled in ice and stirred well during the addition. The resulting green solution was quickly filtered. Solid ethylenediamine-N,n"-diacetic acid (8.89 g, 0.05 mol) was added to the filtrate. The solution was heated a t 50' for 5 min. The resulting purple solution was cooled in ice and three volumes of 95% ethanol was gradually added. A purple oil separated. The supernatant liquid was decanted, the oil was redissolved in a minimum amount of water, and about one volume of 95% ethanol was slowly added. Purple crystals usually separated a t this point. If an oil formed, the above procedure was repeated until crystals were obtained; yield, 5.5 g. The air-dried crystals lost 3 mol of water per mole when dried in a vacuum oven a t 40" for 36 hr. The addition of strong acid to the complex caused evolution of COZ. Resolution.-(-))ars-[Co(en)20x]Br~H20 (7.3 g, 0.02 mol) and silver acetate (3.3 g, 0.02 mol) were combined and stirred for several minutes in 25 ml of water a t 60". The AgBr which had formed was removed and washed with warm water (5 ml). Solid K[trans-Co(EDDA)CO3] .HzO (4 g, 0.012 mol) was added to the filtrate with good stirring. After the solid had dissolved, the solution was cooled in ice for 1-2 hr. The red diastereoisomer (2.6 g) was removed. This material was recrystallized twice from warm water by cooling in ice; yield, 1.5 g, with AsSsj -5.7. Anal. Calcd for [Co(en)lox][Co(EDDA)COa]~3H20: C, 25.7; H , 5.26; N, 13.8. Found: C, 25.69; H , 5.48; N, 13.87. Since optically active trans-Co(EDDA)COa- could not be isolated from the diastereoisomer, its CD spectrum was obtained by subtracting the CD curve of (-)546-Co(en)2oxf from the curve of the diastereoisomer. The validity of this procedure was checked by the comparison of the CD curve of K[trans-(-)Co(EDDA)ox] with the curve obtained by the same procedure [(- )-Co(EDDX)ox] studied over a sixfold for [ ( - )~j-Co(en)20x] concentration range. The curves obtained directly and by subtraction of the contribution for the resolving agent agreed within 5'3, of Ae or better. The differences between the total areas of peaks were insignificant. Optical activity due to the trans-Co(EDDA)COs- was found to decrease rapidly for aqueous solutions of the diastereoisomer. Reproducible CD values were obtained by making measurements on cold solutions of the diastereoisomer immediately after mixing. A solution was replaced by a fresh one when i t no longer gave reproducible data (2-3 min after mixing in ice water). Seven solutions were required to obtain a complete CD curve for the 650-380-mp region of the spectrum. If (+)5~Co(en)2ox+was substituted for the ( - )546 isomer in the resolution procedure, the diastereoisomer containing the (+ ) isomer of trans-Co(EDDA)COs- was obtained. Preparation of Potassium trans-Ethylenediamine-N,N'-diacetato(oxalato)cobaltate(III) Dihydrate, K[trans-[Co(EDDA)ox] .2HzO.-Cobalt(II) carbonate (9.0 g, 0.075 mol) and ethylenediamine-N,N'-diacetic acid (13.2 g, 0.075 mol) were stirred together in 75 ml of water a t 60" for about 20 min. The resulting mixture was added to a solution of K2Cz04 (9.2 g, 0.05 mol) and H&Oa.2H20 (3.2 g, 0.025 mol) in 200 ml of water. Five grams of activated charcoal was added and air was bubbled through the mixture for 17 hr. The charcoal was removed and the red-purple filtrate was evaporated under a stream of air on a steam bath to a volume of 100 ml. The solution was cooled in ice for 2 hr. Purple crystals of K[trans-Co(EDDA)ox].2H10 (3.1 g ) were collected. The filtrate was left standing in the refrigerator for several days and a little ethanol was added to obtain 9 g of red solid. Recrystallization of this solid from warm water by cooling in ice gave 1 g of K[trans-Co(EDDA)ox].2HzO as the least soluble fraction. Anal. Calcd for K[Co(EDDd)ox] .2H20: C, 24.2; H, 3.54; N, 7.08. Found: C, 23.86; H , 3.71; N, 6.93.
Resolution.-(+)546-[Co(en)20x] I (2.3
g, 0.006 mol) and silver
Inorganic Chemistry acetate (0.9 g, 0.0058 mol) were stirred together in water (15 ml) a t 60". After several minutes AgI was removed and washed with water (5 ml). The filtrate was warmed to 60" and solid [trans-Co(EDDd)oxj.2H20 (2.2 g, 0.006 mol) was added with rapid stirring. The red diastereoisomer began t o separate almost a t once. The first fraction was removed after about 2 min. This material (1.3 g, AeSjs +4.6) was recrystallized once from warm water by cooling in ice. Fractions of diastereoisomer with Aeajs >+5.5 (based on molecular weight assuming 1:l diastereoisomer of 588) were combined (0.8 g) and stirred for several minutes with K I (1.5 g) in a few milliliters of water. The resolving agent was removed and K[trans-Co(EDDA)ox].2Hz0 (0.35 g , 4 e ~ f5.2) 0 was recovered from the filtrate by addition of ethanol (several volumes) and cooling in ice. If necessary, the complex can be recrystallized from cold water by adding ethanol and cooling in ice. The K ( - )-[trans-Co(EDDB)ox].2H20 was obtained from the filtrate after removal of the diastereoisomer. This solution was treated with about 2 g of K I and excess resolving agent was removed. Ethanol was then added to precipitate partially resolved K ( - )-[trans-Co(EDDA)ox].2H20. This material was fractionally recrystallized from cold water by adding ethanol and cooling in ice. Fractions (0.35 g) were obtained with Aejso -5.2. Preparation of Potassium trans-Ethylenediamine:N,N'-diacetato(malonato)cobaltate(III)Trihydrate, K [trans-Co(EDDA)mal] 3HzO.-Cobalt(II) carbonate (9.6 g, 0.08 mol) and ethylenediamine-N,N'-diacetic acid (14.0 g, 0.08 mol) were stirred together in 80 ml of water a t 60" for about 20 min. The resulting mixture was added to a solution of malonic acid (10.4 g, 0.1 mol) and KOH (7.6 g, 0.14 mol) in water (80 ml). Three grams of activated charcoal was added and air was bubbled through the mixture overnight (15 hr). The charcoal was removed and the filtrate was evaporated under a stream of air on a steam bath until purple solid began to separate (40-50 ml). The mixture was cooled in ice for 1 hr. Purple K[trans-Co(EDDA)mal] 3He0 was removed and washed with ethanol and acetone and air dried; yield, 7.5 g. Resolution.-(+)S46-[Co(en)~ox]I (2.8 g, 0.007 mol) and silver acetate (1.1g, 0.007 mol) were stirred together for several minutes in water (10 ml) a t 60". The AgI which had separated was removed and washed with 10 ml of water. Solid K[trans-Co(EDDA)mal].3H%O(2.5 g, 0.0067 mol) was added to the cooled filtrate with good stirring. An equal volume (20 ml) of 95% ethanol was added to the resulting solution. Red solid separated a t once and was quickly removed. This material was recrystallized from a minimum amount of cold water by addition of ethanol to give 1.2 g with A e j ~ j-3.65 (assuming anhydrous 1: 1 diastereoisomer). The diastereoisomer was stirred with NaI (3.5 g) in a few milliliters of water for several minutes. After removal of the resolving agent, the malonato complex was precipitated by adding ethanol and cooling in ice. The sodium salt of the complex was isolated because the potassium salt did not crystallize satisfactorily. The fractions obtained were recrystallized from cold water by addition of ethanol to give maAnal. Calcd for Na[Co(EDDA)mal] . terial with A6666 -4.7. 3H20: C, 26.2; H , 4.36; N, 6.80. Found: C, 26.72; H, 4.39; N, 6.87. The filtrate from the diastereoisomer was evaporated under a stream of air a t room temperature to a volume of 30 ml. NaI (5 g ) was added with stirring and after several minutes the resolving agent was removed. The (+) isomer of the malonato complex was obtained by the same procedure as that used for precipitating and purifying the (- ) isomer. Spectra.-The absorption spectra were measured on a Cary Model 14 recording spectrophotometer and the CD curves were recorded with a Roussel- Jouan Dichrograph using a Sylvania Sun Guu light source. The CD spectra were measured in a 2-cm cell a t room temperature with the exception of the diastereoisomers with Co(EDDA)C03- which were measured in ice-cold M. Proton solutions. Concentrations used were about magnetic resonance spectra were obtained on a Varian A-60 spectrometer (60 Mcps). The solvent was DiO; DSS (sodium 9
VoZ.8, No. 1, January
CIRCULAR DICHROISM OF COBALT(III) COMPLEXES117
1969
2,2-dimethyl-2-silapentane-5-sulfonate)was added to serve as an internal standard. Analyses.-Elemental analyses were performed by Alfred Bernhardt, Elbach, West Germany.
Results and Discussion Preparation of a complex thought to be K(transCo(EDDA)COa].HzO has been reported.6 The preparative procedure was the same as that used in this study, but the isolation procedure and the product obtained were different. Several attempts were made to repeat this previous work, but in no case were the red-violet crystals reported obtained. The complex prepared by Mori and coworkers was characterized by its cobalt analysis and absorption spectrum. I t s visible absorption spectrum showed only one maximum (526 mp) and a shoulder (380 mp), whereas the visible spectrum of the carbonato complex isolated in the present study shows two maxima (Figure 1). The great similarity of the visible absorption, circular dichroism, and proton magnetic resonance spectra for the latter complex to those of the similar oxalato and malonato complexes tends to indicate that it is the trans-Co(EDDA)COa- complex. Since the complex reported previously could not be prepared in this laboratory, no speculation as to its identity is possible. 16
18
20
24
22
26
quencies and chemical shifts for these AB patterns are given in Table I. The fact that no cis isomers of these complexes were isolated is consistent with the finding of Legg and Cooke7that the trans configuration of coordinated EDDA is preferred. TABLE I PMRFREQUENCIES FOR AB PATTERN DUE TO ACETATE RINGPROTONS K [Co(EDDA)COs]
K[Co-
Na[Co-
(EDDA)oxl
( E D D A) mal 1
4.37 4.07 3.43 3.13
4.42 4.12 3.45 3.15
4.37 4.07 3.38 3.08
4.14 4.28 3.20 3.33 JABwas 18 cps for all three complexes.
4.20 3.25
Resonance freq from DSS, ppm
Acetate ring proton chem shifts,“ cps
H, Hb a
The shapes of the CD curves for the complexes investigated (Figure 1) are very similar to those for the related isomers of trans-Co(EDDA)en + and transCo(EDDA) (S)-ala (Figure 2).* I n the following discussion trans-Co(EDDA)ox- will be used as a representative complex for the three new complexes reported in this paper. Since the three complexes are very similar, the same arguments should apply to all three.
28
X(rnp) 6 5 0 600
-K(+)-frans- [ C o ( E D D A l o x ] * 2 H z O ---- Na(+l-trans-[Co(EDDA)rnol].
4.80
3H20
550
500
400
(+I- t E - [Co(EDDAl(S-olol]~2H~O
A€
---K(+I---
1
-1 .o
450
[ C o l E D D A I o x ] . ZH,O
A€
~
100 -1.60 6 5 0 600 5 5 0
500
450
400
350
p)
I 16
I 18
I
I
22 20 B (ern-’) x
I
I
I
24
26
28
Figure 1.-Circular dichroism and molar absorptivity vs. wave number for the ($)F,~B isomers of trans-Co(EDDA)CO%-, trans-Co(EDDA)ox-, and trans-Co(EDDA)mal-.
Figure 2.-Circular dichroism vs. wave number for the ( - t ) : ~ g isomers trans-Co(EDDA)en+, trans-Co(EDDA)(S)-ala, and tr~ns-Co(EDDA)ox-.
Proton nuclear magnetic resonance was used to assign the trans configuration to EDDA in the complexes studied. Since the two acetate rings of transcoordinated EDDA are in equivalent environments, only one AB pattern is expected for the two sets of acetate ring protons. For the cis isomer, two different AB patterns would be expected, one for the protons of each acetate ring. The nmr spectra of similar complexes have been thoroughly d i ~ c u s s e d . ~ Each complex in the present series gave one AB quartet due to acetate ring protons. The resonance fre-
If chelate rings are ignored, the effective crystal field symmetry of trans-Co (EDDA)ox- is cis-CoNz04 while that of trans-Co(EDDA)en + is trans-CoOzN4. By the electrostatic model for treating crystal field potentials] both of these complexes (ignoring chelate rings) have tetragonal symmetry. Under tetragonal symmetry the TI, ( o h ) level, which is the first excited state for low-spin Co(III), is split into levels with Az and E symmetry. Whether the E state lies lower than the A2 state or higher depends upon the arrangement of the ligands and their relative crystal field strengths.
(6) M. Mori, M. Shibata, E. Kyumo, and F. Maruyama, Bull. Chem. SOC. Japan, SS, 75 (1962). ( 7 ) J. I. Legg and D. W. Cooke, Inoig. Chem., 4, 1576 (1965).
(8) J. I. Legg, D. W. Cooke, and B. E. Douglas, ibid., 6, 700 (1967). (9) C. J. Ballhausen, “Introduction t o Ligand Field Theory,” McGrawHill Book Co., Inc., New York, N. Y., 1962, pp 106 f.
118 CAROLYN W. VAN SAUNAND BODIEE. DOUGLAS When cis-CoNzOl and trans-CoOZNl complexes are compared, i t is predicted that the E level will lie lower than the A2 level for both complexes. This same order of energy levels for the two types of complexes is predicted because both the arrangement of ligands and their relative crystal field strengths are changed in going from one complex to the other. If only one or the other of these things had been changed, the order of levels would have reversed. Under the actual C2 symmetry of the tmns-EDDA complexes, the E level should split into two nondegenerate levels, one with A and one with B symmetry. For complexes of the type being considered, this latter splitting is apparently small enough so that it does not need to be considered, in general, when using CD curves for relating absolute configurations of the coniplexes. Absolute configurations can be assigned to the isomers of the EDDA complexes prepared for this study by comparison of their CD spectra with that for (+)w trans-Co(EDDA)en+. The absolute configuration of the latter complex (Figure 3) is known from an nmr study of an analogous complex.* Since the new complexes have the same order of energy levels in the region of the TI, ( o h ) level (at least in the tetragonal approximation) and the same arrangement of chelate rings, the isomers with the same absolute configuration as (+)h*&ans-Co(EDDA)en+ are the ones for which the CD band of lowest energy (related to A1 + E) is positive (Figures 2 and 3).
N
(+),,,trans-[Co(EDDA)
en]+
Figure 3.-Absolute
( + ) 5 8 9 z s -[Co(EDDAlox]-
(-)546-[Co (EDTA)]-
configurations of some cobalt(II1) complexes.
The complexes trans-Co(EDDA)ox- and Co (EDTA)are both of the type cis-CoNzOdand would be expected to have the same order of energy levels in the TI, (Oh)
Inorganic Chemistry region. As expected, their CD curves do have the same general shape, although the absolute intensities of the bands differ for the two complexes. Coordinated EDTA contains a backbone structure which can be compared with trans-coordinated EDDA. It is interesting to note, however, that (+):~~-CO(EDTX) -, which has the same configuration of this backbone as does (+)-trans-Co(EDDA4)ox-, gives a CD curve for which the corresponding bands have signs opposite those for (+)-trans-Co(EDDA)ox-.3 It is ( -)w Co(EDTA)- which has the same absolute configurationlo as (+)-trans-Co(EDDA)ox- (Figure 3). This relationship is predicted by any one of the several methoddl which can be used for finding the net helicity due to a given arrangement of chelate rings in a complex. The theoretical basis for this observed relationship betn een the net helicity of a multidentate chelate complex and the sign of the CD band due to the -41 + E transition (or the related transition in complexes of lower symmetry) is not understood. The rotational strengths of the d-d transitions of these new complexes are noteworthy for two reasons. The rotational strengths observed are significantly greater than those for most cobalt(II1) complexes with similar ligands such as EDTA (hexa- or pentadentate), ethylenediamine, oxalate, etc. Also the rotational strength seems to be largely determined by the EDDA and is quite insensitive to changes in the bidentate ligand used. For this reason no significant change in rotational strength is seen as the size of the bidentate chelate ring is varied in the carbonato, oxalato, malonato series. There is also no significant difference in rotational strengths between these complexes and the comparable complexes with ethylenediamine or (5')alanine More experimental and theoretical work on the origin of rotational strength in these systems is needed before this dominating influence of the EDDA ligand can be explained. Acknowledgments.-The authors wish to thank Air. L. R. Froebe for measurement of the nmr spectra used in this study. Support in the form of a hlellon predoctoral fellomhip to C. W. Van Saun is also gratefully acknowledged. (10) An error in Figure 8 of ref 8 should be noted. T h e isomer of Co( E D T A ) - pictured is (+):rs-Co(EDTA)- rather t h a n t h e ( - 1 8 4 6 isomer as labeled. T h e isomers are correctly identified in t h e discussion. (11) 5. I. Legg and B. E. Douglas, J. A m . Chem. Soc., 88, 2696 (1966); C. J. Hawkins and E. Larsen, Acta Chew. Scnnd., 1 9 , 196Y (1965); forthcoming tentative IUPAC rule for absolute configurations.
